Machine_Learning.xml

<?xml version="1.0" encoding="utf-8"?>
<!DOCTYPE nta PUBLIC '-//Uppaal Team//DTD Flat System 1.1//EN' 'http://www.it.uu.se/research/group/darts/uppaal/flat-1_2.dtd'>
<nta>
	<declaration>// Place global declarations here.
chan remove_noise_chan , format_data, k_means, fetch_content, automation_model_chan,reommend;
bool recon_complete = false;
bool automation_model_completed = false;
bool user_profile = false;
bool preprocessing_done = false;
bool data_formatting = false;
bool clustering_completed = false;
bool recon_generated = false;</declaration>
	<template>
		<name x="5" y="5">automation_model</name>
		<parameter>int accuracy</parameter>
		<declaration>bool train_test_data_split = false;


</declaration>
		<location id="id0" x="792" y="-288">
			<name x="782" y="-318">Exit</name>
		</location>
		<location id="id1" x="640" y="-288">
			<name x="616" y="-320">Deploy</name>
		</location>
		<location id="id2" x="480" y="-288">
			<name x="432" y="-320">Check_accuracy</name>
		</location>
		<location id="id3" x="352" y="-288">
			<name x="304" y="-320">Make_predictions</name>
		</location>
		<location id="id4" x="208" y="-288">
			<name x="176" y="-272">Train_Model</name>
		</location>
		<location id="id5" x="24" y="-288">
			<name x="-16" y="-272">Model_Selection</name>
		</location>
		<location id="id6" x="-240" y="-288">
			<name x="-288" y="-320">Feature_Engineering</name>
		</location>
		<location id="id7" x="-712" y="-288">
			<name x="-816" y="-320">Data_Gathering</name>
		</location>
		<location id="id8" x="-488" y="-288">
			<name x="-552" y="-328">Data_Preprocessing</name>
		</location>
		<location id="id9" x="493" y="-34">
			<name x="450" y="-17">Got_recommendations</name>
		</location>
		<init ref="id7"/>
		<transition>
			<source ref="id9"/>
			<target ref="id1"/>
			<label kind="guard" x="561" y="-153">recon_generated == true</label>
		</transition>
		<transition>
			<source ref="id2"/>
			<target ref="id9"/>
			<label kind="guard" x="391" y="-170">accuracy &lt; 90</label>
			<label kind="synchronisation" x="399" y="-187">reommend!</label>
		</transition>
		<transition>
			<source ref="id5"/>
			<target ref="id5"/>
			<label kind="guard" x="-80" y="-432">train_test_data_split == false</label>
			<label kind="assignment" x="-88" y="-456">train_test_data_split = true</label>
			<nail x="-56" y="-384"/>
			<nail x="80" y="-384"/>
		</transition>
		<transition>
			<source ref="id4"/>
			<target ref="id4"/>
			<label kind="guard" x="153" y="-442">clustering_completed == false</label>
			<label kind="synchronisation" x="176" y="-432">k_means!</label>
			<nail x="136" y="-408"/>
			<nail x="272" y="-408"/>
		</transition>
		<transition>
			<source ref="id1"/>
			<target ref="id0"/>
			<label kind="assignment" x="592" y="-264">automation_model_completed = true</label>
		</transition>
		<transition>
			<source ref="id2"/>
			<target ref="id1"/>
			<label kind="guard" x="512" y="-288">accuracy &gt;= 90</label>
		</transition>
		<transition>
			<source ref="id3"/>
			<target ref="id2"/>
			<label kind="assignment" x="348" y="-263">accuracy = accuracy+5</label>
		</transition>
		<transition>
			<source ref="id4"/>
			<target ref="id3"/>
			<label kind="guard" x="221" y="-238">clustering_completed == true</label>
		</transition>
		<transition>
			<source ref="id5"/>
			<target ref="id4"/>
			<label kind="guard" x="34" y="-323">train_test_data_split == true</label>
		</transition>
		<transition>
			<source ref="id6"/>
			<target ref="id5"/>
		</transition>
		<transition>
			<source ref="id8"/>
			<target ref="id8"/>
			<label kind="guard" x="-646" y="-459">data_formatting == true</label>
			<label kind="synchronisation" x="-648" y="-480">remove_noise_chan!</label>
			<nail x="-688" y="-456"/>
			<nail x="-512" y="-456"/>
		</transition>
		<transition>
			<source ref="id8"/>
			<target ref="id8"/>
			<label kind="guard" x="-739" y="-119">data_formatting == false</label>
			<label kind="synchronisation" x="-720" y="-136">format_data!</label>
			<nail x="-776" y="-136"/>
			<nail x="-600" y="-136"/>
		</transition>
		<transition>
			<source ref="id8"/>
			<target ref="id6"/>
			<label kind="guard" x="-467" y="-289">preprocessing_done == true</label>
		</transition>
		<transition>
			<source ref="id7"/>
			<target ref="id8"/>
			<label kind="synchronisation" x="-688" y="-312">automation_model_chan?</label>
		</transition>
	</template>
	<template>
		<name>remove_noise</name>
		<declaration>bool remove_na_null = false;
bool outliers_removal = false;</declaration>
		<location id="id10" x="25" y="-59">
			<name x="15" y="-89">Noise_removed</name>
		</location>
		<location id="id11" x="-365" y="-59">
			<name x="-383" y="-97">detect_outliers</name>
		</location>
		<location id="id12" x="-672" y="-56">
			<name x="-720" y="-48">Fill_missing_values</name>
		</location>
		<location id="id13" x="-904" y="-56">
		</location>
		<init ref="id13"/>
		<transition>
			<source ref="id11"/>
			<target ref="id11"/>
			<label kind="guard" x="-408" y="76">outliers_removal == false</label>
			<label kind="assignment" x="-403" y="53">outliers_removal = true</label>
			<nail x="-421" y="53"/>
			<nail x="-293" y="53"/>
		</transition>
		<transition>
			<source ref="id12"/>
			<target ref="id12"/>
			<label kind="guard" x="-739" y="-212">remove_na_null == false</label>
			<label kind="assignment" x="-739" y="-229">remove_na_null = true</label>
			<nail x="-736" y="-184"/>
			<nail x="-608" y="-184"/>
		</transition>
		<transition>
			<source ref="id11"/>
			<target ref="id10"/>
			<label kind="guard" x="-229" y="-85">outliers_removal == true</label>
			<label kind="assignment" x="-238" y="-59">preprocessing_done = true</label>
		</transition>
		<transition>
			<source ref="id12"/>
			<target ref="id11"/>
			<label kind="guard" x="-586" y="-85">remove_na_null == true</label>
		</transition>
		<transition>
			<source ref="id13"/>
			<target ref="id12"/>
			<label kind="synchronisation" x="-840" y="-80">remove_noise_chan?</label>
		</transition>
	</template>
	<template>
		<name>data_cleansing</name>
		<location id="id14" x="-51" y="-59">
		</location>
		<location id="id15" x="-272" y="-56">
		</location>
		<location id="id16" x="-480" y="-56">
		</location>
		<init ref="id16"/>
		<transition>
			<source ref="id15"/>
			<target ref="id14"/>
			<label kind="guard" x="-238" y="-85">data_formatting == false</label>
			<label kind="assignment" x="-238" y="-59">data_formatting = true</label>
		</transition>
		<transition>
			<source ref="id16"/>
			<target ref="id15"/>
			<label kind="synchronisation" x="-408" y="-80">format_data?</label>
		</transition>
	</template>
	<template>
		<name>k_means_algo</name>
		<parameter>int cluster_size</parameter>
		<declaration>int counter = 0;</declaration>
		<location id="id17" x="314" y="-51">
			<name x="304" y="-85">exit</name>
		</location>
		<location id="id18" x="-102" y="-51">
			<name x="-142" y="-43">assign_points</name>
		</location>
		<location id="id19" x="-357" y="-51">
			<name x="-405" y="-91">random_sample</name>
		</location>
		<location id="id20" x="-654" y="-51">
			<name x="-664" y="-81">start</name>
		</location>
		<init ref="id20"/>
		<transition>
			<source ref="id18"/>
			<target ref="id17"/>
			<label kind="guard" x="42" y="-76">counter &gt;= cluster_size</label>
			<label kind="assignment" x="17" y="-51">clustering_completed = true</label>
		</transition>
		<transition>
			<source ref="id19"/>
			<target ref="id18"/>
			<label kind="guard" x="-255" y="-76">counter == 0</label>
		</transition>
		<transition>
			<source ref="id20"/>
			<target ref="id19"/>
			<label kind="synchronisation" x="-535" y="-76">k_means?</label>
		</transition>
		<transition>
			<source ref="id18"/>
			<target ref="id18"/>
			<label kind="guard" x="-153" y="-153">counter &lt; cluster_size</label>
			<label kind="assignment" x="-144" y="-170">counter = counter + 1</label>
			<nail x="-14" y="-131"/>
			<nail x="-153" y="-127"/>
		</transition>
	</template>
	<template>
		<name>front_end</name>
		<parameter>bool existing_user, bool manual_search</parameter>
		<declaration>bool logged_in = false;
bool create_profile = false;
bool content_access = false;</declaration>
		<location id="id21" x="739" y="-51">
			<name x="729" y="-81">logout</name>
		</location>
		<location id="id22" x="629" y="-51">
			<name x="619" y="-81">watch</name>
		</location>
		<location id="id23" x="168" y="-192">
			<name x="158" y="-222">click_content</name>
		</location>
		<location id="id24" x="-433" y="-85">
			<name x="-465" y="-69">login</name>
		</location>
		<location id="id25" x="408" y="-56">
			<name x="365" y="-102">check_subscription</name>
		</location>
		<location id="id26" x="168" y="16">
			<name x="136" y="24">search_content</name>
		</location>
		<location id="id27" x="-170" y="-85">
			<name x="-263" y="-59">recommend_content_generated</name>
		</location>
		<location id="id28" x="-484" y="110">
			<name x="-540" y="134">profile_creation_process</name>
		</location>
		<location id="id29" x="-773" y="-85">
			<name x="-783" y="-115">Start</name>
		</location>
		<init ref="id29"/>
		<transition>
			<source ref="id29"/>
			<target ref="id24"/>
			<label kind="guard" x="-637" y="-85">existing_user == true</label>
		</transition>
		<transition>
			<source ref="id28"/>
			<target ref="id29"/>
			<label kind="assignment" x="-739" y="110">existing_user = true</label>
			<nail x="-773" y="110"/>
		</transition>
		<transition>
			<source ref="id25"/>
			<target ref="id25"/>
			<label kind="guard" x="323" y="93">content_access == false</label>
			<label kind="assignment" x="323" y="76">content_access = true</label>
			<nail x="320" y="64"/>
			<nail x="484" y="68"/>
			<nail x="480" y="56"/>
		</transition>
		<transition>
			<source ref="id22"/>
			<target ref="id21"/>
		</transition>
		<transition>
			<source ref="id25"/>
			<target ref="id22"/>
			<label kind="guard" x="459" y="-42">content_access == true</label>
		</transition>
		<transition>
			<source ref="id23"/>
			<target ref="id25"/>
		</transition>
		<transition>
			<source ref="id26"/>
			<target ref="id23"/>
		</transition>
		<transition>
			<source ref="id27"/>
			<target ref="id23"/>
			<label kind="guard" x="-85" y="-204">manual_search == false &amp;
recon_complete == true</label>
		</transition>
		<transition>
			<source ref="id27"/>
			<target ref="id26"/>
			<label kind="guard" x="0" y="-48">manual_search == true</label>
		</transition>
		<transition>
			<source ref="id24"/>
			<target ref="id27"/>
			<label kind="synchronisation" x="-357" y="-85">fetch_content!</label>
			<label kind="assignment" x="-365" y="-110">logged_in = true</label>
		</transition>
		<transition>
			<source ref="id29"/>
			<target ref="id28"/>
			<label kind="guard" x="-595" y="8">existing_user == false</label>
		</transition>
	</template>
	<template>
		<name>fetch_content_process</name>
		<location id="id30" x="144" y="-88">
		</location>
		<location id="id31" x="-32" y="-88">
		</location>
		<location id="id32" x="-304" y="-88">
		</location>
		<location id="id33" x="-552" y="-88">
		</location>
		<init ref="id33"/>
		<transition>
			<source ref="id31"/>
			<target ref="id30"/>
			<label kind="guard" x="-24" y="-48">automation_model_completed == true</label>
			<label kind="assignment" x="-16" y="-72">recon_complete = true</label>
		</transition>
		<transition>
			<source ref="id32"/>
			<target ref="id31"/>
			<label kind="synchronisation" x="-232" y="-112">automation_model_chan!</label>
		</transition>
		<transition>
			<source ref="id33"/>
			<target ref="id32"/>
			<label kind="synchronisation" x="-480" y="-112">fetch_content?</label>
		</transition>
	</template>
	<template>
		<name>recommendation_system</name>
		<declaration>bool item_based_completed = false;
bool user_based_completed = false;
bool content_based_completed = false;</declaration>
		<location id="id34" x="374" y="-85">
			<name x="364" y="-119">end_result</name>
		</location>
		<location id="id35" x="-170" y="0">
			<name x="-180" y="-34">user_based</name>
		</location>
		<location id="id36" x="-612" y="127">
			<name x="-663" y="144">content_based</name>
		</location>
		<location id="id37" x="-170" y="-221">
			<name x="-180" y="-255">Item_based</name>
		</location>
		<location id="id38" x="-459" y="-85">
			<name x="-486" y="-133">Consolidate</name>
		</location>
		<location id="id39" x="-1045" y="-85">
			<name x="-1109" y="-125">Get_user_profile</name>
		</location>
		<location id="id40" x="-289" y="-340">
			<name x="-299" y="-374">top_10_item</name>
		</location>
		<location id="id41" x="-238" y="110">
			<name x="-255" y="127">top_10_users</name>
		</location>
		<location id="id42" x="-782" y="0">
			<name x="-816" y="25">top_10_content</name>
		</location>
		<init ref="id39"/>
		<transition>
			<source ref="id41"/>
			<target ref="id38"/>
			<label kind="assignment" x="-374" y="42">user_based_completed = true</label>
		</transition>
		<transition>
			<source ref="id35"/>
			<target ref="id41"/>
		</transition>
		<transition>
			<source ref="id42"/>
			<target ref="id38"/>
			<label kind="assignment" x="-799" y="-68">content_based_completed = true</label>
		</transition>
		<transition>
			<source ref="id36"/>
			<target ref="id42"/>
		</transition>
		<transition>
			<source ref="id40"/>
			<target ref="id38"/>
			<label kind="assignment" x="-527" y="-280">item_based_completed = true</label>
		</transition>
		<transition>
			<source ref="id37"/>
			<target ref="id40"/>
		</transition>
		<transition>
			<source ref="id38"/>
			<target ref="id37"/>
			<label kind="guard" x="-306" y="-187">recon_generated == false &amp; 
item_based_completed == false</label>
		</transition>
		<transition>
			<source ref="id38"/>
			<target ref="id34"/>
			<label kind="guard" x="68" y="-153">item_based_completed == true &amp;
user_based_completed == true &amp;
content_based_completed == true</label>
			<label kind="assignment" x="102" y="-85">recon_generated = true</label>
		</transition>
		<transition>
			<source ref="id38"/>
			<target ref="id35"/>
			<label kind="guard" x="-348" y="-68">item_based_completed == true &amp;
user_based_completed == false</label>
		</transition>
		<transition>
			<source ref="id38"/>
			<target ref="id36"/>
			<label kind="guard" x="-612" y="8">user_based_completed == true
&amp; content_based_completed == false</label>
		</transition>
		<transition>
			<source ref="id39"/>
			<target ref="id38"/>
			<label kind="synchronisation" x="-833" y="-110">reommend?</label>
		</transition>
	</template>
	<system>// Place template instantiations here.
ml_model = automation_model(80);
rm_noise = remove_noise();
k = k_means_algo(4);
data_clean = data_cleansing();
f = front_end(false, false);
fetch_content_obj = fetch_content_process();
recon_system = recommendation_system();

// List one or more processes to be composed into a system.
system f, ml_model, recon_system, rm_noise, data_cleansing, k, fetch_content_obj ;</system>
	<queries>
		<query>
			<formula>A[] deadlock imply f.logout</formula>
			<comment>No deadlock : A[] deadlock imply f.exit</comment>
		</query>
		<query>
			<formula>A[] not( ml_model.accuracy &gt;= 90 &amp;  recon_generated )</formula>
			<comment>It should not happen that a model gave an accuracy of 90% and also the recommendations are generated.</comment>
		</query>
		<query>
			<formula>A[] k.counter &lt;= k.cluster_size</formula>
			<comment>Cluster size for K means : A[] k.counter &lt;= k.cluster_size</comment>
		</query>
		<query>
			<formula>A[] (f.watch imply ( f.existing_user and f.content_access) )</formula>
			<comment>If watch then it must be an existing user and must have the subscription</comment>
		</query>
		<query>
			<formula>A[] ( f.click_content imply ( f.existing_user &amp; f.logged_in &amp; (recon_complete | f.manual_search) ) )</formula>
			<comment>Only be able to click content if he is a existing user and logged in and either got the recommemdations or manually searched.</comment>
		</query>
		<query>
			<formula>A[] not( ml_model.accuracy &gt;= 90 &amp;  recon_generated )</formula>
			<comment>I don't want a stage where I got accuracy &gt;= 90 and also running the recommendation system.</comment>
		</query>
		<query>
			<formula>A[] ( f.logout imply !(f.logged_in == false | f.existing_user == false ) )</formula>
			<comment>I dont want a stage where the user is going to logout but the logged_in status is false and existing_user status is also false.</comment>
		</query>
		<query>
			<formula></formula>
			<comment></comment>
		</query>
		<query>
			<formula>A&lt;&gt; (f.login imply f.recommend_content_generated)</formula>
			<comment></comment>
		</query>
		<query>
			<formula>A&lt;&gt; (f.click_content imply f.check_subscription)</formula>
			<comment>If content clicked then check for subscription : A&lt;&gt; (f.click_content imply f.check_subscription)</comment>
		</query>
		<query>
			<formula>A&lt;&gt; (f.recommend_content_generated imply ml_model.Exit)</formula>
			<comment>If i reach recommendation then ml should exit : A&lt;&gt; (f.recommend_content imply ml_model.Exit)</comment>
		</query>
		<query>
			<formula>A[] ( ml_model.Deploy imply ( ml_model.accuracy &gt;= 90 | recon_generated == true ) )</formula>
			<comment>Deploying only successfull model : A[] ( ml_model.Deploy imply ( ml_model.accuracy &gt;= 90 | recon_generated == true ) )</comment>
		</query>
		<query>
			<formula>A[] ( ml_model.Model_Selection imply ( preprocessing_done == true ) )</formula>
			<comment>Only use model after Preprocessing is done.</comment>
		</query>
		<query>
			<formula>A[] ( f.profile_creation_process imply (!f.existing_user)  )</formula>
			<comment>Create profile only if user is not and existing user else Just Login.</comment>
		</query>
		<query>
			<formula>A[] ( ml_model.Make_predictions imply ( preprocessing_done &amp; clustering_completed ) ) </formula>
			<comment>Model should make prediction only when preprocessing is completed and k means clustering is completed.</comment>
		</query>
		<query>
			<formula>A[] ( ml_model.Exit imply automation_model_completed )</formula>
			<comment>Machine learning model should only exit after the automation model has completed its execution.</comment>
		</query>
		<query>
			<formula>A[] ( rm_noise.Noise_removed imply (rm_noise.remove_na_null &amp; rm_noise.outliers_removal )  )</formula>
			<comment>Noise Removal model should only exit after it has removed null values and also removed the outliers.</comment>
		</query>
		<query>
			<formula>A[] ( ((k.counter &lt; k.cluster_size)  and clustering_completed == false) or (clustering_completed imply (k.counter &gt;= k.cluster_size) )  )</formula>
			<comment>If the counter is less than the cluster size I should not have clustering_completed as true. Also, if cluster_completed is true then counter should be greater than cluster size.</comment>
		</query>
		<query>
			<formula></formula>
			<comment></comment>
		</query>
		<query>
			<formula>E&lt;&gt; f.watch</formula>
			<comment>User will always reach watch state.</comment>
		</query>
		<query>
			<formula>E&lt;&gt; f.click_content</formula>
			<comment>Always exist a path where we will reach click_content</comment>
		</query>
		<query>
			<formula>E&lt;&gt; ( f.click_content | f.search_content )</formula>
			<comment>There always exist a path where the user can click the recommended content or can search for content.</comment>
		</query>
		<query>
			<formula>E&lt;&gt; ( ml_model.Deploy )</formula>
			<comment>There always exist a stage in future where we will be deploying model.</comment>
		</query>
		<query>
			<formula>E&lt;&gt; ( ml_model.Model_Selection ) </formula>
			<comment>There always exist a state where we will be having model selection.</comment>
		</query>
		<query>
			<formula>E&lt;&gt; ( recon_system.end_result )</formula>
			<comment>The recommendor system should produce the output results at some stage.</comment>
		</query>
		<query>
			<formula>E&lt;&gt; f.logout</formula>
			<comment>The user should be able to navigate to logout page.</comment>
		</query>
		<query>
			<formula>E&lt;&gt; f.login</formula>
			<comment>The user should be able to navigate to login page.</comment>
		</query>
	</queries>
</nta>